Examensarbeten för masterexamen // Master Theses
Länka till denna samling:
Browse
Browsar Examensarbeten för masterexamen // Master Theses efter Program "Engineering mathematics and computational science (MPENM), MSc"
Visar 1 - 20 av 179
Sökresultat per sida
Sortera efter
- PostA Dispatching Algorithm with Application to Fleets of Shared Autonomous Vehicles(2017) Hellsten, Erik; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical SciencesFleets of shared autonomous vehicles have been predicted to dominate the transport sector within a near future. For this to work efficiently—including the handling of spontaneous requests—the associated routing problems need to be modelled dynamically and solved efficiently. We formulate and model the problem of routing a fleet of shared autonomous vehicles over a period of time. For each vehicle and each moment in time, it must be decided which customers to serve and which routes to take. The resulting model is solved using a rolling horizon optimisation methodology together with an insertion heuristic for new requests. The optimisation problems resulting from the rolling horizon methodology are solved using column generation, where the subproblems, being elementary shortest path problems with side constraints, are solved using both a local-search heuristic and a dynamic programming algorithm. Our computational experiments show that real-world sized problem instances can be solved to near-optimality within a reasonable computing time.
- PostA generalized vehicle routing problem with spatial and temporal synchronization. Mathematical modelling and solution.(2016) Mustedanaganic, Amir; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical Sciences
- PostA Model for Object Recognition in Liver Resection Surgery(2020) AL-MALEH, CHRISTIAN; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Lundh, Torbjörn; Lundh, TorbjörnLaparoscopic liver resection is a safer alternative to open surgery for treatment of liver cancer, a disease which claims almost 800 000 lives every year. The procedure involves making small incisions in the abdomen where instruments and a camera, called a laparoscope, are inserted. One of the major drawbacks of laparoscopic surgery is the restricted view and orientation, as well as lack of haptic feedback. Incorporating Augmented Reality, or AR, in the laparoscopic view is a proposed method of facilitating the navigation. This work extends a previous model for projecting information from 3D to 2D and vice versa using reference points, which correctly visualizes the shape, angle and size of a tumor in AR in the 2D laparoscopic view. To enable the 2D-to-3D projection, two object recognition models based on image segmentation and edge detection, respectively, were developed where white reference objects were distinguished from the darker tones of the liver tissue. The positions of the reference objects were then measured. The latter model, albeit effective given certain frames, failed to identify fiducials over the course of a test film. Since the process of image segmentation is computationally heavy, it was localized to an area of interest in a given frame, reducing the algorithm’s runtime. Statistical error estimation was used to validate the positions found by this recognition model. The average position error produced was between 1 to 5 pixels, where the frames had a pixel height of 1080. Future work involves combining the recognition algorithm with the projection model to examine the effect of the deviations of the estimated positions in the 2D laparoscopic view.
- PostA space-time cut finite element method for a time-dependent parabolic model problem(2015) Lundholm, Carl; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical SciencesIn this thesis, a space-time finite element method for the heat equation on overlapping meshes is presented. Here, overlapping meshes means that we have a stationary mesh of the solution domain with an additional mesh that is allowed to move around in and through the solution domain. The thesis contains a derivation, an analysis, and results from an implementation of the method. The derivation starts with a strong formulation of the problem and ends with a finite element variational formulation together with adequate function spaces. For the finite element solution, we use continuous Galerkin in space and discontinuous Galerkin in time, with the addition of a discontinuity in the solution on the space-time boundary between the two meshes. In the analysis, we propose an a priori error estimate for the method with discontinuous Galerkin of order zero and one, i.e., dG(0) and dG(1). For dG(1), the error estimate indicates that the movement of the additional mesh decreases the order of convergence of the error, with respect to the time step, from the third to the second order, when the speed of the moving mesh is large enough. The order of convergence with respect to the step size for dG(1), as well as the error convergence for dG(0), are unaffected by the moving mesh and are thus as in the case with only a stationary mesh, presented in [2, 3]. An implementation of the method in one spatial dimension, with piecewise linear elements in space, and dG(0) and dG(1) in time, has also been performed. The numerical results of the implementation show the superiority of using dG(1) instead of dG(0) for overlapping meshes. The numerical results also confirm the behaviour of the error convergence, indicated by the a priori error estimate. Keywords: partial differential equation, finite element method, space-time cut, time-dependent, parabolic problem, heat equation, overlapping mesh, moving mesh, discontinuous Galerkin, a priori.
- PostA study of the stratification of plane cubic curves and its various generalizations(2013) Boutry, Pierre; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical Sciences
- PostAbstractive Document Summarisation using Generative Adversarial Networks(2018) Svensson, Karl; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical SciencesThe use of automatically generated summaries for long texts is commonly used in digital services. In this thesis, one method for such document summarisation is created by combining existing techniques for abstractive document summarization with LeakGAN – a successful approach at text generation using generative adversarial networks (GAN). The resulting model is tested on two different datasets originating from conventional newspapers and the world’s largest online community: Reddit. The datasets are examined and several important differences are highlighted. The evaluations show that the summaries generated by the model do not correlate with the corresponding documents. Possible reasons are discussed and several suggestions for future research are presented.
- PostActive Learning and Malware Entity Extraction. A survey of active learning methods specifically implemented with a CRF for finding malware names - The hunt for Red October.(2016) Romare, Elin; Bijelovic, Milica; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical Sciences
- PostActive Safety for car-to-bicyclist accidents(2014) Ranjbar, Arian; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical SciencesDuring the last years the rapid development of countermeasures has led to an overall decreasing number of fatalities in road traffic accidents. However, the positive trend does not hold at the same rate for bicyclists. It is therefore desireable to investigate these type of accidents and discuss possible countermeasures. In this work the car-to-bicyclist accidents were studied to understand the context in which they occur. Data was collected from the German In-Depth Accident Study (GIDAS) database and the Pre-Crash Matrix (PCM) extension which contained reconstructed accidents. A geometrical classification method was developed to find the most common type of accident scenarios and descriptive statistics was gathered. Further on, a risk model was derived to find the probability of an accident to result in severe injuries. The model was also used to calculate the theoretical effectiveness of a simplified Autonomous Emergency Braking (AEB) system. The study showed that the most common car-to-bicyclist accidents were lateral and longitudinal scenarios in intersections, where the car was travelling straight forward. This was also connected to the risk analysis which showed that the most risk in uencing parameter was the impact speed of the car in the collision. Finally, the effectiveness study indicated that AEB has a considerable potential to mitigate car-to-bicyclist accidents.
- PostAdjustable Percolation(2014) Stolze, Sebastian; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical Sciences
- PostAdversarial Representation Learning for Synthetic Replacement of Sensitive Speech Data(2020) Östberg, Adam; Ericsson, David; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Mostad, Petter; Listo Zec, EdvinAs more data is collected in various settings across organizations, companies, and countries, there has been an increase in the demand of user privacy. Developing privacy preserving methods for data analytics is thus an important area of research. In this work we present a model based on generative adversarial networks (GANs) that learns to obfuscate specific sensitive attributes in speech data. We train a model that learns to hide sensitive information in the data, while preserving the meaning in the utterance. The model is trained in two steps: first to filter sensitive information in the spectrogram domain, and then to generate new and private information independent of the filtered one. The model is based on a CNN that takes mel-spectrograms as input. A MelGAN is used to invert the spectrograms back to raw audio waveforms. We show that it is possible to hide sensitive information such as gender by generating new data, trained adversarially to maintain utility and realism.
- PostAge and age Distribution Estimation from Images and video(2016) Säbben, Olivia; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical Sciences
- PostAI Enabled Service Market Logistics(2020) Ramne, Johan; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Lang, Annika; Lang, AnnikaUncertainty about upstream suppliers’ ability to deliver ordered quantities on time is one the reasons that manufacturers and retailers need to keep safety stock in inventory. Through accurate prediction of suppliers’ delivery performance the uncertainty can be quantified and used by material planners in their decision-making process. Representing the deliver performance of an individual supplier as a time series, the uncertainty can be predicted through probabilistic forecasting: estimation of the future probability distribution given past observations. This thesis presents two recurrent neural network models, using encoder-decoder architectures, for multi-step ahead probabilistic forecasting of the delivery performance of suppliers to Volvo Group Trucks Operations Service Market Logistics. The models are evaluated on mean quantile loss for a number of quantiles over a 14 week forecast range. One model, DeepAR, outperformed exponential smoothing models generated by the forecast package in R on four out of five quantiles.
- PostAI-Based Toxicity Prediction as an Alternative to Animal Testing(2023) Dalman, Mercedes; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Kristiansson, Erik; Kristiansson, Erik; Gustavsson, MikaelIn recent years, there has been a significant increase in the use of chemicals in our environment due to growing demand and consumption. Consequently, large-scale chemical regulation based on toxicological assays has been implemented to prevent exposure-related consequences for nature and human health. Historically, animal-based assays have been used for this purpose. However, there is now an increasing demand to replace these animal-based assessment methods with computer-based alternatives. Despite previous attempts to develop computer-based models, these models have proven to be unreliable and inaccurate, leading to a decrease in interest. Therefore, there is a pressing need to develop new computer-based models for toxicity assessment. Here, the introduction of deep learning models, particularly transformer architecture, has the potential to revolutionise the field. Deep neural networks have demonstrated the ability to handle complex and high-dimensional problems, surpassing older modelling techniques. Moreover, as the transformer has shown promise in handling chemical structure information, there is growing interest in its usage in the field of environmental toxicity assessment. The aim of this project was hence to explore the potential of transformer-based deep neural network models for the purpose of toxicity assessment. For this project, a subset of rat and mice in vivo toxicity assay data associated with EC50 and LOEC measurements, as well as different administration routes, were utilised. Here, three sets of data were analysed, each distinguished by the hazards: acute toxicity, carcinogenicity, or reproductive toxicity. The first type of model, the single-DNN model, was created for each data set separately. Subsequently, these models were expanded to the multiple-DNN model, able to handle all three data sets simultaneously. For all models, a pre-trained RoBERTa transformer was utilised to interpret canonicalised SMILES representation of chemical structures, with the performance then evaluated through repeated 10-fold crossvalidation. Principal Component Analysis demonstrated that the transformer could identify patterns in chemical structures related to toxicity. Moreover, the study found that the single-DNN model outperformed the multiple-DNN model in all trials, likely due to the latter’s increased complexity. All models exhibited leniency towards chemicals with low measured concentrations, and to mitigate this problem, a more stringent loss for lower concentrations was suggested. Overall, this project demonstrated the potential and effectiveness of transformer-based computer models for toxicity assessment, showcasing the versatility of this technology for addressing a broad range of toxic hazards
- PostAlgorithms for Robust Path-Planning(2020) Jonsson Damgaard, Thomas; Åkerlund, Jens; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Andersson, Claes; Andersson, Claes; Averö, Anders; Ellrén, Patrik; Rittri, Mikael; Warston, HåkanRoute optimization is a commonly studied field of optimization resulting in pathplanning algorithms. In this project, alternative route generation and robustness analysis were conducted for common road networks and off-road terrains. This was done using Open Street Map data and high-resolution terrain data provided by Vricon and Lantmäteriet. Alternative routes were generated using a constructed 2-way search algorithm. Robustness analysis was split into a physical robustnessindex and a non-physical robustness simulation. The generated routes conform to the constructed robustness-index, which is sensitive to weather conditions when traversing terrain. The simulation is applied to the generated routes to visualize additional penalty times where the user can identify critical points along said route. This is done using a robustness simulation which creates obstacles along the chosen path and calculates the additional time needed. The final result is a complete Java program connected to Carmenta Engine, a map engine provided by Carmenta. The resulting program showed conceptually promising routes, providing reasonable alternatives while accounting for current weather conditions.
- PostAlgorithms for Robust Path-Planning(2020) Jonsson Damgaard, Thomas; Åkerlund, Jens; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Andersson, Claes; Andersson, Claes; Averö, Anders; Ellrén, Patrik; Rittri, Mikael; Warston, HåkanRoute optimization is a commonly studied field of optimization resulting in pathplanning algorithms. In this project, alternative route generation and robustness analysis were conducted for common road networks and off-road terrains. This was done using Open Street Map data and high-resolution terrain data provided by Vricon and Lantmäteriet. Alternative routes were generated using a constructed 2-way search algorithm. Robustness analysis was split into a physical robustnessindex and a non-physical robustness simulation. The generated routes conform to the constructed robustness-index, which is sensitive to weather conditions when traversing terrain. The simulation is applied to the generated routes to visualize additional penalty times where the user can identify critical points along said route. This is done using a robustness simulation which creates obstacles along the chosen path and calculates the additional time needed. The final result is a complete Java program connected to Carmenta Engine, a map engine provided by Carmenta. The resulting program showed conceptually promising routes, providing reasonable alternatives while accounting for current weather conditions.
- PostAlternative Pricing in Column Generation for Airline Crew Rostering(2018) Curry, Emily; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical SciencesIn airline crew rostering, the objective is to create personalized schedules, i.e., rosters, for a set of crew members. Because of the large number of possible rosters that could be formed, the problem is solved using column generation, where each column corresponds to a specific roster. The pricing problem, which is the problem studied in this thesis, is then defined as to find legal rosters with the potential of improving the current solution. Since the rules and regulations regarding rosters vary between airlines, we have chosen to treat the pricing problem as a black-box optimization problem. Three different methods for solving the black-box pricing problem have been implemented. The first method uses binary particle swarm optimization (BPSO) to search for new rosters. The other two methods use surrogate modeling to fit a nonlinear surrogate function to a set of sampled rosters using radial basis functions. The surrogate function was then either linearly approximated, so that a shortest path problem could be set up and solved, or solved heuristically by a BPSO method. The three methods have been evaluated on five real-world test cases. For each test case, a large number of different pricing problems are solved. Our comparison of the methods’ performance shows that the method using BPSO performed the best, followed by the surrogate modeling approach without the linear approximation.
- PostAn Implementation of the Branch-And-Price Algorithm Applied to Opportunistic Maintenance Planning(2015) Friberg, David; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical Sciences
- PostAn Improved Shortest Path Algorithm for Aircraft Routing(2022) Thorén, Samuel; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Strömberg, Ann-Brith; Grönkvist, MattiasThousands of commercial flights are scheduled each day, and the need for flight routes to be as efficient as possible is very important. The cost for the flights has to be as low as possible while still fulfilling and complying with all regulations in safety and other requirements. Aircraft routing is the problem of creating routes for aircraft to operate, subject to constraints. The shortest path problem with resource constraints is commonly used as a subproblem in solution methodologies in order to solve the aircraft routing problem. The aim of this thesis is to implement and evaluate an algorithm called a multidirectional dynamic programming algorithm (MDDPA) in order to solve the shortest path problem with resource constraints. The results from the implementation of MDDPA is compared to the results of another dynamic programming algorithm. The results show that MDDPA in most test cases finds a solution with a cost that is at least as good as the solution found by the other dynamic programming algorithm. However, this usually comes at the expense of longer execution times, where our implementation of MDDPA in most cases takes longer time at finding a solution. In order to gain better results in terms of computing times, we suggest implementing multithreading in MDDPA and analyse how the choice of parameters used in MDDPA could affect the results.
- PostAnalysing a modified ranking algorithm for exploratory search(2020) Fällman, Markus; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Sagitov, Serik; Jonasson, JohanExploratory Search is a small emerging field within Information Retrieval, studying a type of searching called exploratory searching. This type of search is directed towards learning and investigating, and has recently started to draw attention. However, the field of Exploratory Search struggles with its methodology. A central problem is the difficulty to measure improvements due to that exploratory searching by definition lacks precise goals. New tools and ideas are therefore often evaluated with user studies. By focusing on describing how tools and ideas work, researchers can avoid the difficulty and contribute to the field. Such an indirect approach allows formulating measures that can be applied to ranked lists, which, in turn, allow using simulations with many benefits. This study showcases the approach. The aim is to determine if a ranking algorithm modification influence the formation of groups in lists of ranked articles returned from an academic search engine. The data are generated by simulated searches and a Linear Mixed Model is used for the analysis. The main covariates represent how the ranking of a standard ranking algorithm is weighted together with the ranking according to two new criteria. The response variable consists of scores on how tightly connected the ranked articles are, with the importance of links decreasing with the depth, and comes from the application of a measure developed in the thesis. The main result is that the level of interconnectedness between high ranking articles can be clearly and statistically significantly influenced by the modification, although the influence varies with the randomly generated queries. While more research is needed, this might be useful for controlling the articles interconnectedness when constructing a search engine. On a different level, the thesis shows how the indirect approach can be applied, that it enables using simulations, and it indicates that the approach can produce results interesting for exploratory searching.
- PostAnalysis of gene expression patterns in wild eelpout from the Baltic Sea - Identification of expression patterns for differentially expressed genes connected to pollution(2014) Wijkmark, Emma; Chalmers tekniska högskola / Institutionen för matematiska vetenskaper; Chalmers University of Technology / Department of Mathematical Sciences